axis-java-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Ruchith Fernando <ruchith.ferna...@gmail.com>
Subject [Axis2] Re: Getting there from here (Re: [jira] Commented: (AXIS2-359) Stack overflow when processing XSD with mutual recursion)
Date Fri, 06 Jan 2006 11:19:47 GMT
Adding the [Axis2] prefix  just in case someone missed the mail

On 1/6/06, Davanum Srinivas <davanum@gmail.com> wrote:
> Team,
>
> Q: How many test cases are in Axis 1.X's tree that run w2j and run
> junit test cases using the generated code?
> A: 110
>
> Q: How many tests are in Axis2's tree that run w2j?
> A: 12
>
> Q: Is there a test for testing imports (wsdl/xsd)?
> A: No
>
> Q: Were there problems reported before for imports?
> A: Yes, http://issues.apache.org/jira/browse/AXIS2-337.
>
> Q: How many tests use the Axis2's w2j generated code and run the junit test?
> A: 1 (AFAICT just the security test, i am not sure if the interop
> tests are even hooked up to "all-tests", need to check)
>
> Q: Should we support null's in the generated code?
> A: I'd think yes and ASAP! What do you think?
> http://issues.apache.org/jira/browse/AXIS2-260
> http://issues.apache.org/jira/browse/AXIS2-321
>
> Q: Should we handle exceptions/faults in wsdl correctly?
> A: I'd think yes and ASAP! What do you think?
> http://issues.apache.org/jira/browse/AXIS2-360
>
> Q: Does one need deep knowledge of Axis 1.X to do the "right" thing
> for example with this one?
> (http://marc.theaimsgroup.com/?l=axis-user&m=113640313810607&w=2)
> A: No. One just needs to run WSDL2Java in Axis 1.X and look at the
> generated package names.
>
> Q: Do we have a problem with not trying to understand user needs?
> A: I think we may have a slight problem, because in our quest for
> Technical superiority/perfection, we may be losing a bit of focus on
> the customers. heck, we were not even picking up a custom wsdl for a
> codegen generated service from the aar file till yesterday
> (http://issues.apache.org/jira/browse/AXIS2-350). This has nothing to
> do with Axis 1.X. We need to test our code better and watch each
> others backs better.
>
> Q: Do i agree that what we have in Axis2 is the best architecture?
> A: Absolutely!!!! Yes, but when we implement something, we need to
> check Axis 1.X behavior first and question why something was done
> (even just looking at the artifacts will help!)
>
> Q: Can we do better?
> A: Absolutely!!!! Yes, we can.
>
> Q: Do we have the *BEST* team in the world?
> A: Yep, we do!!!!
>
> Q: Is dims cranky all the time?
> A: Am i? :)
>
> Q: Is sanjiva defensive all the time?
> A: Is he? :)
>
> Paul,
> We need more people looking at Axis2 with a critical eye, from the
> point of view of end users w/o regard to how swamped people are or why
> a certain task is not yet done or why a specific corner was cut. Could
> you help?
>
> thanks,
> -- dims
>
> On 1/5/06, Paul Fremantle <pzfreo@gmail.com> wrote:
> > Actually I think getting the balance right from Axis1 is the right thing to
> > do. I was just using Axis1 WSDL2Java and I don't think its by any means
> > perfect. I would say that ADB and WSDL2Java is probably the one area of
> > Axis2 that needs the most work, but I also think its an area where we can
> > easily improve over Axis1.
> >
> > I also think its an area where we can make fast progress. It seems to me
> > that the new architecture can be much more effective than the old
> > architecture.
> >
> > Dims - maybe it would be useful if you or one of the old hands in Axis1
> > could go through old JIRAs or your deep memory banks and come up with a list
> > of potential issues you think we should deal with, and then we can work
> > through them. The fact is that there is no simple way we can "learn" from
> > Axis1 other than the considered and helpful input of the guys who did it
> > first time.
> >
> > The only other thing is that WSDL2Java v2 is a complete mystery to me - and
> > maybe to others. Perhaps we need to make a concerted effort to get one or
> > two more people up and running on it so that we can progress faster.
> >
> > Paul
> >
> > PS I don't think its a mess ..... its just at an earlier stage than the
> > other parts of Axis2. I've been playing with WSDL2Java quite a bit. I'm not
> > sure it helps to say its a mess when people are working hard at improving
> > it!
> >
> > PPS One area that needs some improvement is getting line numbers in error
> > messages - or we need a WSDL validator to run first. I tend to have to use
> > Axis1 to debug my WSDL before I run Axis2.
> >
> >
> >  On 1/5/06, Sanjiva Weerawarana (JIRA) <jira@apache.org> wrote:
> > >     [
> > http://issues.apache.org/jira/browse/AXIS2-359?page=comments#action_12361868
> > ]
> > >
> > > Sanjiva Weerawarana commented on AXIS2-359:
> > > -------------------------------------------
> > >
> > > Um, is the fix to this just the stuff that was in the original issue
> > report? If so that hardly qualifies as  "deep mess".
> > >
> > > Come on Dims, we *do* leverage Axis1 .. I can't believe you are saying
> > that we should get much/anything out of the old symboltable mess that
> > Axis1's WSDL2Java was/is. Just because there are certain cases we don't
> > handle doesn't mean we are not learning from the past. That's exactly why
> > we're still in 0.93+ and not in 1.xx.
> > >
> > > BTW
> > http://soapinterop.java.sun.com/round3/groupd/imported/import2B.wsdl
> > does not appear to be WS-BP 1.0 compliant; did they not outlaw importing
> > XSDs using wsdl:import?
> > >
> > > > Stack overflow when processing XSD with mutual recursion
> > > >
> > --------------------------------------------------------
> > > >
> > > >          Key: AXIS2-359
> > > >          URL:
> > http://issues.apache.org/jira/browse/AXIS2-359
> > > >      Project: Apache Axis 2.0 (Axis2)
> > > >         Type: Bug
> > > >   Components: wsdl
> > > >     Versions: 0.9, 0.91, 0.92, 0.93
> > > >  Environment: Win2K, Java   1.4.2_08-b03, Axis2 0.91, Axis2 0.93
> > > >     Reporter: Kevin J. Winters
> > >
> > > >
> > > > The recursive method WSDLPump.pushSchemaElement(...) does not protect
> > against schemas with mutual recursion.  A check of already processed
> > namespaces could be used to prevent the stack overflow in a manner similar
> > to:
> > > >    ...
> > > > /** begin mod **/
> > > >     private HashSet schemaCache = new HashSet();
> > > > /**  end mod  **/
> > > >
> > > >     private void pushSchemaElement(Schema originalSchema,Stack stack){
> > > >         if (originalSchema==null){
> > > >             return;
> > > >         }
> > > >
> > > > /** begin mod **/
> > > >         if (
> > schemaCache.contains(originalSchema.getDocumentBaseURI()))
> > {
> > > >             return;
> > > >         } else {
> > > >
> > schemaCache.add(originalSchema.getDocumentBaseURI());
> > > >         }
> > > > /**  end mod  **/
> > > >         stack.push(originalSchema);
> > > >         ...
> > >
> > > --
> > > This message is automatically generated by JIRA.
> > > -
> > > If you think it was sent incorrectly contact one of the administrators:
> > >
> > http://issues.apache.org/jira/secure/Administrators.jspa
> > > -
> > > For more information on JIRA, see:
> > >    http://www.atlassian.com/software/jira
> > >
> > >
> >
> >
> >
> > --
> > Paul Fremantle
> > VP/Technology, WSO2 and OASIS WS-RX TC Co-chair
> >
> > http://bloglines.com/blog/paulfremantle
> > paul@wso2.com
> >
> > "Oxygenating the Web Service Platform", www.wso2.com
> >
>
>
> --
> Davanum Srinivas : http://wso2.com/blogs/
>


--
Ruchith

Mime
View raw message